Objective: This study investigates speech characteristics associated with physiological and pathological aging, using Parkinson’s disease (PD) as a model, with the aim of identifying linguistic and acoustic digital biomarkers that differentiate these groups.

Background: Previous research suggests that digital speech biomarkers may provide a non-invasive, scalable, and objective tool for assessing cognitive decline in neurodegenerative conditions (Kovac et al., 2025). The present study examines how aging affects speech patterns.

Method: The dataset included 35 healthy young subjects (HY; mean age 24.6 ± 4.4), 37 healthy aged subjects (HA; mean age 68.2 ± 4.5), and 32 subjects with PD (mean age 70.8 ± 6.9). We analyzed recordings and automatic transcriptions from three speech tasks (~90 s each): spontaneous monologue, reading, and the Cookie Theft Picture Description (CTPD). Speech was evaluated using natural language processing and acoustic analysis. Group differences were tested using Mann–Whitney U and Kruskal–Wallis tests, with associations assessed using Spearman correlation.

Results: Compared with HY speakers, HA speakers showed lower speech rate with fewer words and syllables per minute, longer pauses, and more frequent and longer breathing events, producing shorter speech segments between breaths. Linguistically, their speech was slightly less elaborated and more disfluent, with more filler expressions. The PD group differed significantly from both other groups. Acoustically, PD speakers showed reduced articulatory range and more monotonous speech. Linguistically, their speech was shorter and syntactically simpler, with fewer words per sentence, reduced use of several grammatical categories, and more interjections and fillers. PD speakers also used the past tense less frequently and the present tense more often, suggesting a preference for descriptive rather than narrative discourse. They also showed lower speech rate, longer and more frequent pauses, increased breathing events, and shorter speech segments between breaths.

Conclusion: Physiological aging mainly affected speech rate, pausing, and breathing behavior. Compared with both intact groups, PD patients exhibited reduced articulatory variability, monopitch, lower speech rate, increased pausing and respiratory activity, and a simplified linguistic structure.

References: Kovac et al. https://doi.org/10.1016/j.compbiomed.2025.111251

« Back to 2026 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/digital-speech-biomarkers-reveal-speech-changes-in-physiological-aging-and-parkinsons-disease/